Sign in with
Sign up | Sign in
Your question
Solved

Are long stress tests necessary?

Tags:
  • Overclocking
  • Stress Test
  • Processors
Last response: in Overclocking
Share
May 17, 2014 5:18:50 PM

Doesn't your processor reach its max temp in minutes when under load? Why do you have to do it for several hours or more? Is ok to run it for only 30 minutes or so when adjusting your OC? So, raise multiplier until it crashes, then raise voltage etc. until you reach target OC, THEN stress test for a few hours?

More about : long stress tests

May 17, 2014 6:10:53 PM

From my experience the best Stress Test is to leave it for a couple of weeks at your prefered oc!! Play your games normal etc... watch for your temperatures and such with a monitoring programm. See for any freezes or unnatural behavior everything is fine then you are fine!!!

It is usually good to use prime and other stress tests, ( Do not use intel burn test, it kills everything. I do not see the point to stress it that much when all your apps will never even stressed it so much) to see the temps right away!! Cinebench is good and fast way to see temps and check for any throttling and see how effective your oc is!!
m
0
l
May 17, 2014 6:16:04 PM

I read that memory plays a role in overclocking as well. Like when you OC your CPU, you have to change some RAM settings as well. Is that just because they want to OC their RAM or does CPU and RAM OC'ing go hand in hand? In other words can I just OC my CPU and leave ALL RAM settings alone? I don't really care about RAM very much.
m
0
l
Related resources
May 17, 2014 6:21:44 PM

Well in all the Unlocked Processors ( Unlocked are the ones where you can change the Multiplier) which includes the Intel k series and all the Amd, there is no need to change anyhting in your ram. In locked cpu's you have to change the FSB where it has impact in Cpu clock speed as well as Ram.
m
0
l
May 17, 2014 6:37:31 PM

OK I see, so the ONLY settings I have to tamper with are voltage and the multiplier then since I have a K version? No other setting in my BIOS?
m
0
l
a b K Overclocking
May 17, 2014 8:50:32 PM

Yes long stress testing for your 24/7 overclock is important if you value system stability. If it doesn't pass a true stress test for at least 3 hours it's not stable period.
m
0
l
May 17, 2014 9:05:52 PM

Do all tests push your cpu to 100% and just keep it there? What makes some tests (prime 95, Intel burn etc) "harder" than others and what's the difference between a game than maxes your cores vs. a test if they both bring your cpu to 100% anyway?
m
0
l
a b K Overclocking
May 17, 2014 9:16:00 PM

a stress test such a p95 will put an artificial 100% load on every core the CPU has. You will most likely never see a load as harsh as this (closest I have found that approaches p95 heat levels is handbrake encoding 1080p at custom settings, still about 6-7c lower than p95) but for an overclock to be stable it has to be able to run at full throttle without crashing or throwing errors.
m
0
l
May 18, 2014 6:42:15 AM

Ok, I ask because BF4 maxes out my cores so I guess I'm asking if 100% from a game vs. 100% from a stress test are different in "toughness" or if 100% is the same no matter how it happens (gaming or stress testing). Basically, can 100% usage be more demanding with different tests?
m
0
l
a b K Overclocking
May 18, 2014 7:30:08 AM

Yes 100% games isn't as demanding as p95. Prime95 is about as hard as you can push a cpu.
m
0
l
May 18, 2014 7:42:47 AM

As long as I make sure my temps are ok, there's no worry of prime95 "killing" my cpu even if I ran if for a week or something insane?
m
0
l
a b K Overclocking
May 18, 2014 7:47:06 AM

As long as your temperatures are fine then no damage will occur. Running a heavy stress test for that long would be to no benefit. Do keep in mind how much power you may be using overclocked and the limits of your power supply. If it passes 6-8 hours of p95 blend test with no errors or dropped cores, then I would consider that overclock stable for 24/7 everyday use.
m
0
l
May 18, 2014 7:53:32 AM

lastly, if you're in the process of manual overclocking, and you are changing settings then testing then changing settings etc. Can you run a stress test for a short amount of time (like 30 mins) before increase the ratio (frequency)? Then when you're happy, do a final long test (6-8 hours) to really test it, or do you have to do really long tests every time you make a small change (like 100Mhz increase). If doing it that long way it would take days to get from 3.4Ghz to say 4.2Ghz.
m
0
l

Best solution

a b K Overclocking
May 18, 2014 8:15:51 AM

I do a quick 30min run myself when playing around with frequencies and voltages to check quickly for stability.
Share
May 18, 2014 8:29:53 AM

Ok thanks for the help.
m
0
l
!